Datasets:

sil-ai

/

bloom-vist

like 3

Follow

SIL Global - AI 43

Stock-taking:

Items that need updating:

• Searching for "todo" in the README shows Dataset Description section is empty
• Preprocessing field also
• Data Fields also
• We get warnings about the YAML format that "licenses" is deprecated. Use "license" instead.
• "languages" is deprecated. Use "language" instead.
• speaking of licenses, we need to add the extra gated prompt.

What have we got:

• There's a pretty good explanation of the VIST-format JSON.
• explained the language-specific subsets.
• details about parsing, filtering, original data.

extra_gated_prompt: |-
  One more step before getting this model.
  This model is open access and available to all, with a CreativeML OpenRAIL-M license further specifying rights and usage.
  The CreativeML OpenRAIL License specifies: 

  1. You can't use the model to deliberately produce nor share illegal or harmful outputs or content 
  2. CompVis claims no rights on the outputs you generate, you are free to use them and are accountable for their use which must not go against the provisions set in the license
  3. You may re-distribute the weights and use the model commercially and/or as a service. If you do, please be aware you have to include the same use restrictions as the ones in the license and share a copy of the CreativeML OpenRAIL-M to all your users (please read the license entirely and carefully)
  Please read the full license here: https://huggingface.co/spaces/CompVis/stable-diffusion-license
  
  By clicking on "Access repository" below, you accept that your *contact information* (email address and username) can be shared with the model authors as well.
    
extra_gated_fields:
 I have read the License and agree with its terms: checkbox

Comparing with bloom speech, here are the sections that exist

Dataset Summary
Languages
Dataset Statistics
Dataset Structure
    Data Instances
    Data Fields
    Data Splits
Changelog

...whereas we have these:

Dataset Description
    Downloading Annotations
    Dataset Preprocessing
Dataset Structure
    Data Instances
    Data Fields
    Data Splits
    Internal VIST-Format JSON
    Subsets
    Languages
Dataset Creation
    Filtering and Curation

Oh, and according to the card we don't have specified splits either.

There's also the fact that the format of the JSON is mimicking the original VIST SIS, but then this dataset of ours doesn't really, so it's a pretty arbitrary format.

OK, updated the YAML at least.

All right, as for the Dataset Description, how would I want someone to explain this to me?

I think we've got a number of updates we need to make to the dataset itself. For one thing, it's still pulling from https://bloom-vist.s3.amazonaws.com/bloom-vist.json, an outdated version of the JSON. There's also some issues with the data format. Really, if people want to be using this they should pull from the deduped language filtered JSON directly. https://bloom-vist.s3.amazonaws.com/bloom_vist_june15_deduped_langfiltered.json

We also don't have assigned splits.

I think I'm just going to throw all these caveats into the README

Updated the README to some extent.

Can't solve #2, #3 and #4 in time for EMNLP, so we're going to revert to baaa8e0ade0f1fe5d49ff99af8cd492f5d6f80e9 for now and add some notes.

So I will try this:
https://stackoverflow.com/a/21718540

Even this did not work. I am getting missing album errors after trying to revert to the commit I thought would be working.